Discussion in 'Miami Dolphins Forum' started by Jerrysanders, Jul 4, 2017.
Thanks for the kind gesture but analytics are self-explanatory.
On the same hand, if quoting posts are too difficult to figure out, I can also offer you help.
Yeah, I am a believer in analytics. But like anything else, it's garbage in = garbage out. There is an old saying, "lies ... damned lies ... and statistics." That is because misused or improperly balanced analytics can paint a very incorrect picture. It's kinda like me putting on an official NFL jersey and thinking I'm a bona fide NFL player ... it's a farce. Hopefully one day this kid learns how to use analytics and do a proper analysis.
You called numbers useless obviously there not self-explanatory in your case; actually based on the arguments of certain posters on this forum they also don't get analytics.
Um Digital your arguments about offensive Line play, sack total over 4 years, coaching and lack of TE plays implies you don't get the analytics he used.
Concerning football and specifically in this case they're useless when used alone, useless without context, and useless when unsupported by film.
We get it. We really do. Most fully understand what the analytics have concluded here but it still doesn't mean it's a sound argument being presented. Breaking down football and/or it's players cannot be done thru numbers alone.
For those of us who cannot see the now deleted quote, here is the information of the post it was quoting:
Jerrysanders, post: 7410091, member: 128559
I'm going to assume you wrote something. Based on what Travis34 said, you clearly tried to respond, but either you thought better of what you wrote or the moderators saw fit to delete it for being a personal attack or something. Regardless, we disagree on this video and our assessment of Tannehill.
IF you are going to do any sort of analytics on a QB, it is beyond comprehension that you fail to stress the change in coaching staff and offensive scheme. Tannehill has had 4 sets of coaches in 4 years, with very little continuity. But, it is the current coaching staff that merits the most close scrutiny. With any transition you will see an initial learning period, followed by more stable data once the team has adjusted to the new offense. The Author ignores this. He lumps ALL of Tannehill's 2016 stats into one undivided bucket. Anybody knows that is a poor analysis. The more prudent thing to do, and the way any NFL analyst (and I mean analyst, not reporter or blogger) would do it is to look at the break points.
With Tannehill, we see a break point after 5 games. He threw very few INT's after that point, and his metrics, even by the Author's standards, were stellar. The next break point, obviously, would be the injury which ended his season. But yeah ... why note that, because it would show clearly that once the Dolphins had gotten a handle on the offense that the QB position was performing at a very very high level. If Tannehill plays the 2017 season at a pace similar to how he played the 2016 season after the first 5 games ... we are looking at a lock to make the Pro Bowl. I'm expecting him to be close to that good. It will be interesting to revisit this Author's thoughts at that point.
Actually they can be used alone depending on what question your asking; If that question is has Tannehill been an elite QB, than analytics can determine that without looking at any game film if its good analytical data. Trying to determine how well Tannehill reads defenses or what plays he executes the best those are questions game film is better equipped to answer.
FWIW, his response appears in the last sentence of your post that he erroneously quoted. It's this:
Who the **** around here is saying Tannehill is an elite QB? And why the hell would you need analytics to answer such a rudimentary question?
LOL. Yes, I do understand the analysis. It's an attempt to lengthen the data through a three tiered analysis that ignores many significant contributing factors in an attempt to streamline an elongated process.
This is what is called "little data" and he is trying to say that normalized over time there is some magical QB talent that wins out. Sometimes that's true, but not all the time. When you attempt to normalize an offense with multiple All-Pros and numerous Pro Bowlers vs. an offense that has zero All-Pros and only 1 Pro Bowler ... it renders the Author's point obviously skewed.
Do you honestly think that I do not understand his analysis? Is that even a real question?
I think it's really funny when somebody posts a thread like this and has no problem shoving their viewpoints down somebody else's throat but when the shovel is pointed the other direction, they get all high and mighty.
1. Analytics are useless without context. If I tell you a story about Fred, but you don't know who the **** Fred is, there is no context. Or in football terms, a stat sheet shows a player fumbled. You say to yourself, "hold onto the ****ing ball, dude!" But if you look into the context, you'd see that the center stepped on the QBs foot and the handoff to the RB was muddied.
2. I feel like PFF has breeded a culture of super nerds who like to hide behind stats until somebody fashions the stats in a certain way to refute the original argument. Which then goes back to point 1 of context.
3. Measure, in analytics, the general developmental sluggishness of Tannehill under "mastermind" Joe Philbin. Or the general psychological affect of being sacked 184 times. Or for further CONTEXT, change the variable of the defense being in the top 10. How many more opportunities does the offense have and thus, how much are the numbers accelerated?
In the end, Jerry, refer to the below photo and have a nice day.
Plenty of PPL on this forum believe Tannehill is an elite QB if they thought he was an average or slighly above average QB then there wouldn't be so much backlash every time someone mentions any flaw he has. Analytics is not subject to bias opinions the way game film analysis is, so it might be the best way to answer such a rudimentary question.
Amen. That and the misinformation which comes from those trying to oversimplify the game with box score conclusions (aka fantasy football analysis)
No sir, there are not plenty. There's barely a few, if any at all. That's also quite an agenda a gap in classifications going from average, and slightly above average, to elite.
Analytics and it's processes are very much subject bias. Many conclusions based on numbers alone are riddled with paradoxes and other errors. Game film analysis is not some wishy-washy process when you understand what you are watching and looking for. Now projecting talent, etc. from tape is laden with opinions and bias, but film doesn't lie.
Well you definitely sound like you know more about data analysis then the average person but this is little Data and Big Data ( its better if you have both) but when you mention offensive line and TE play, you have to understand how all of that becomes irrelevant with such a Large sample size on QBs, that means QBs who's had bad wrs,OL play,bad coaches,bad defenses, trouble with the law, still all meet the threshold he's looking for ; with all that being said he's still was unwilling to discount Tannehill even though his NFL career has just been average and is reserving judgement for the 2017. (he may be optimistic because of his college production).
Ex If you tell me Eli Apple has All Pro potential based on the game tape and I say I didn't watch his tape but every single CB in the history of the NFL has an arm length of 32 inches(this is a real stat) and Mr Apple only has a 31 inch arm length. My analysis came from the statistical history of the NFL something my bias can't effect but you judging how well he jams a wr or how smooth his back peddle is 100% subject to bias. Yea analytics in general is subject to bias but this is analytics based on every single QB going back to 1969 its either true or untrue.
Then it should be easy to find posts by people claiming he is elite, right? I'll hang up and wait.....
Do us all a favor and look up the posters that you say that he's elite, then PM them so that you could have this private conversation amongst yourselves. So posters like myself who think Tannehill is just above average with an elite arm don't have to listen to this drivel.
Yep, just like starting threads on Finheaven on the 4th of July.
Don't hold your breath lol
"Elite qbs produce"
Drew Brees? Last 5 years? Why has there been no mention of this? He's an 'elite qb' why haven't the saints been a top level team as of late?
That's the best example there is.
Also...There is only one single "elite" qb in the NFL right now that came out of the shoot producing and that's Rogers. They all took roughly 5-7 years before they consistently put up numbers that they do. Compare RT to ANY playoff QB over the first 4 years of their career and will be right there if not ahead. And NOT ONE of them had to deal with anything close to what he has.
It's beyond ridiculous how we still continue to deal with these agenda driven threads.
This is the purest statement from any post on the subject. A players football iq, speed of the game as well as the "why" something happened all comes from game film. Analytics does nothing in that regard. Watch some tape on RT and you will see all you need to..... if you know what the hell you're seeing unfold. The kid has it between the ears and the arm to compete with ANY qb in the league. That tape don't lie.
Archie Manning was probably the 2nd best Manning but he was mired in mediocrity for the Saints. In his prime on a contending team, he well could have been lights out.
Conversely, QBs like Trent Dilfer and Doug Williams.. and maybe even Jim Plunkett has a .500 winning record, a 52% career completion percentage and 34 more career ints than TDs yet he because of the team he played on, he's been to 10 playoff games.
Sure QBs make a difference, but as Brees and Rivers presently underscore, the team on which they play is more of a determinant of them being succesful.
Don't forget about Phillip Rivers who many believe to be a first ballot hall of famer (pretty much the definition of elite). Why haven't the chargers made any noise if elite qb talent elavates others to win games?
Edit: Vaark beat me to it
Just going to put this right here. The article is geared towards scouting but the principles remain true for subjects like the OP is tossing out in this thread.
Lies, damned lies, and statistics.
This is not how f'ing analytics works!
Data analysis only provides clues or red flags that signal to the researchers that further investigation (other sources of data, observation, process review, etc) is needed. Analytics should NEVER be used to formulate a conclusion. The only way that analytics can lead to a direct conclusion is if the concerned parties already have a pre-conceived agenda. This is where analytics fails, and where those who, like the "source" and the OP, fail at their understanding of the data.
I review data all day. I have one key source of data that I mine every day that tells me people's performance based on a percentage. However, once I see a low performance percentage, I have combine that data with other data, I have to observe the individual, research machines for faults, review the processes being utilized for inconsistencies, and interview his boss for feedback on the individual. Then and ONLY then, can I make a conclusion on the individual.
If I instead rely on only one piece of analytical data, good people would lose there jobs, even when there might be multiple sources that say otherwise.
This "source" and the OP have cherry picked a single source of data which has fit a pre-conceived assumption/agenda, and hid behind their agenda in the name of “analytics." They have both failed miserably to understand what analytics truly is, and the dangers of doing so. By ignoring the basic tenants of analysis and claiming otherwise, they have exposed that they are, in fact, driving a pre-established agenda.
And to make matters worse... We are now discussing "elite" - a subjective term. So, we are using faulty analytics to prove/disprove a subjective categorization?
(edited for typos)
Why have I thought about this guy for the first time in forever ??
My question to the OP is what are your intentions here? Always posting anti Tannehill stuff. Do you think you are going to sway votes in favor of your dislike for the guy simply because you bombard the board with anti Tannehill material? What you are doing is crushing your reputation on this message board and, I don't know if you're familiar with message boards, that doesn't usually end well for said person. They either just disappear or often get so consumed with wanting to prove their once sided point that they break forum rules and are suspended. Why don't you try posting some pro Miami Dolphins material for once?
This is the second most true thing i've read in this thread.
I still think this guy is sourwrong, but he might be Jerry Springer.
OK Jerry considering you are all about those stats, here are 4 QBs, and how they looked in their 5th year, compared to Tannehill. Tell me would you have taken any of those other 4 QB after seeing how they looked in their 5th year?
QB #3- 16...16...288.....474...60.8...3,692...7.8.....28...14.......92.6
QB #4- 16...16...392....591...66.3...4,200...7.1......27...19........88.8
That up there is Ryan's 13 game averaged out stats in 16 games, but considering Ryan was getting very hot at that point before going down, I think I can say those 16 game stats up there are actually very conservative, considering what he might have had.
It's not. Jerry has a better eye than sowrong ever will.
Actually it never becomes irrelevant. The "large" sample size of which you're suggesting would be a playing career of around 200 years. That's when all of those things could fully normalize. But when we're talking about taking an NFL career of 4 years and trying to elongate that through a collegiate career of 2 years and not collecting the data to elongate it through the high school years like the Author assumes is needed for his analysis ... it is the very definition of not enough data. This is the classic "too small" sample size. Clearly 4 years is not going to show any normalization of variances in crucial factors such as offensive line protection, presence of a run game, and average time to throw. It's like comparing 1st and 10 performance with 3rd and long performance ... it's apples to oranges.
I don't want to put words in awsi's mouth, but I do agree with his approach of looking at a Qb's YPA to determine how effective a QB and system are. and this last season when tannehill was clicking it was borderline elite, and as high as it had been for his career (from what I remember...)
You quoted me, so I'll respond.
I am also a big fan of YPA if you are going to use a single stat. The problem is that the Author's analysis didn't use a single stat, he used a half dozen stats but crafted a very narrow swath of information, then tried to extrapolate from there. We see the failing. Of course, I would never spend time watching a 25 minute video on a player's YPA, when I could just look it up. When he professes a full analytics, and only brings this to the table ... it is extremely disappointing.
And as you mentioned Wildbill, when you look at YPA, Tannehill looked stellar. When you look at the stable data, after the initial 5 game learning curve, it looks extremely good. But the Author didn't even bother noting the new offense, learning curve, or adjustment period.
Thanks Hoops, for some reason we have ppl on here stating I'm anti Tannehill by pointing out flaws in his stats , but Ignore the fact that I post more pro Tannehill articles than any of them.
Its much better to use A/YA (better indicator of QB performance)
I'm Sorry But what is that. air yard average?